AI-based AD(Anomaly Detection) for predictive maintenance of

pharmaceutical manufacturing equipment

This is Impix's AI solution for predictive maintenance of manufacturing equipment using AI.

1. Pain Point 

Industrial characteristics

- In order to enhance the global competitiveness of the pharmaceutical industry, it is necessary to secure high competitiveness in production and quality control in terms of the pharmaceutical production base, which is the foundation of overseas exports.

- The company, which operates equipment for producing ampoules, belongs to the typical machinery industry sector. Even a single equipment failure could cause critical issues in mass production, making it an appropriate candidate for predictive maintenance.  

- To achieve this, the manufacturing process must evolve beyond the level of an automated and efficient smart factory to an intelligent factory equipped with a data-driven predictive and preemptive detection system.

Data and System Aspects

- Manufacturing and process data are being accumulated, but there is a lack of direction and methodology on how to utilize the stored data.

- The introduction of an AI system capable of analyzing data is necessary to improve production efficiency and optimize process operations.

- Additionally, the introduction of a system that enables real-time operational visibility and intelligent decision-making support through the automatic collection and integration of real-time field data is required.

Equipment and Infrastructure Aspects

- Methods for constructing equipment utility datasets to prevent deviation from standards are required for equipment operation and maintenance activities.

- The introduction of AI solutions that autonomously detect and predict standard deviation patterns through real-time equipment anomaly detection monitoring is required.

2. AI Solution 

AI AD(anomaly detection) Quality

3. Construction Goals

Detailed Objectives

- Leap beyond the advancement of pharmaceutical factories to intelligent factories.

- Build a system for preemptive detection and prediction of anomalies based on AI data.

- Build a rapid decision-making system through AI-based field data monitoring.

- Apply a predictive maintenance solution to ampoule washers and ampoule fillers on the ampoule manufacturing line.

4. Construction Details

Application of Data Integration Management Technology

- Application of integrated management technology that can simultaneously control data exchange and analysis conditions between heterogeneous equipment.

Application of AI Algorithm Technology

- Detection of abnormal conditions using data extracted from equipment.

- Design of learning models, algorithm extraction, and verification using refined extracted data.

- Application of Anomaly Detection AI algorithm technology through pattern extraction for each piece of equipment.

Predictive maintenance of equipment

- Detect anomalies based on wireless vibration sensors and predict various problems occurring in equipment to prepare for sudden equipment failures in advance.

- By performing predictive maintenance of equipment, large-scale equipment failures can be prevented in advance.

- In addition, the correlation between equipment failures and product quality is tracked and utilized in conjunction with AI big data analysis.

5. Construction Effects

Increased equipment operating rate

- Before adopting: 90% 

- After adopting: 97.5% 

Reduced process defect rate

- Before adopting: 1.5% 

- After adopting: 0.95%